MCP Gateway

Name: MCP Gateway
Author: IBM

@IBM

MCP Gateway について

An AI Gateway, registry, and proxy that sits in front of any MCP, A2A, or REST/gRPC APIs, exposing a unified endpoint with centralized discovery, guardrails and management. Optimizes Agent & Tool calling, and supports plugins.

基本情報

カテゴリ

その他

ライセンス

Apache-2.0

ランタイム

python

トランスポート

stdio

公開者

IBM

設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "mcp-context-forge": {
      "command": "uvx",
      "args": [
        "--from",
        "mcp-contextforge-gateway",
        "mcpgateway",
        "--host",
        "0.0.0.0",
        "--port",
        "4444"
      ]
    }
  }
}

ツール

ツールは検出されませんでした

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is MCP Gateway?

MCP Gateway is an open source registry and proxy that federates Model Context Protocol (MCP) servers, A2A agents, and REST/gRPC APIs into a single unified endpoint. It provides centralized governance, discovery, observability, and protocol translation for AI clients.

How to use MCP Gateway?

Install the mcp-contextforge-gateway package via PyPI or use the Docker image. Configure settings through a .env file or environment variables, then launch the server with the mcpgateway command. A quick start using uvx is also supported.

Key features of MCP Gateway

Federation across MCP, A2A, REST, and gRPC services
REST-to-MCP and gRPC-to-MCP automatic tool translation
Built-in auth, rate-limiting, retries, and OAuth token support
Admin UI for real-time configuration and log monitoring
OpenTelemetry observability with Phoenix, Jaeger, Zipkin, and other backends
Redis-backed caching and multi-cluster federation for scalability

Use cases of MCP Gateway

Federating multiple MCP servers under a single gateway for AI agents
Exposing legacy REST or gRPC APIs as MCP-compliant tools without code changes
Routing agent-to-agent (A2A) requests across OpenAI, Anthropic, or custom agents
Centralizing governance, auth, and observability for enterprise AI infrastructure
Airgapped deployments with a bundled Admin UI and offline capability

FAQ from MCP Gateway

What are the prerequisites to run MCP Gateway?

Python ≥ 3.11 is required. For quick-start smoke tests, curl and jq are recommended.

How do I authenticate with the gateway?

MCP Gateway supports Basic auth, JWT tokens, and OAuth tokens scoped per user. An X-Upstream-Authorization header can be set unconditionally for backend services.

What transports does MCP Gateway support?

It supports HTTP, JSON-RPC, WebSocket, SSE (with configurable keepalive), stdio, and streamable-HTTP.

Does MCP Gateway support OpenTelemetry?

Yes. It provides vendor-agnostic tracing via OTLP protocol and works with Phoenix, Jaeger, Zipkin, Tempo, DataDog, and New Relic backends.

Where is data stored?

MCP Gateway uses Redis for caching and federation. There is no embedded database; configuration and state are managed via files, environment variables, and Redis.

「その他」の他のコンテンツ

Mcp

browsermcp

Browser MCP is a Model Context Provider (MCP) server that allows AI applications to control your browser

6.8K

Activepieces

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

23.2K